GHC proposals not rendering

Simon Peyton Jones

11 Jan 2024 11 Jan '24

10:28 a.m.

If you go to https://github.com/ghc-proposals/ghc-proposals/pull/448, and click on the "rendered" link, namely https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0448-ty..., Github is now displaying the .rst source, not the rendered proposal. This has started happening (consistently across other proposals) today. Does anyone know why? How can I fix this? Simon

Attachments:

attachment.html (text/html — 1.1 KB)

Show replies by date

FWIW, I’m able to render/preview successfully other proposals, but 448-type-variable-scoping loads very slowly and ultimately fails to render for me too. A comment in https://github.com/github/markup/issues/1016 (issue related to .rst not rendering) says that forcing a change in the file triggered a re-render successfully. Rodrigo

...

On 11 Jan 2024, at 15:28, Simon Peyton Jones wrote:

If you go to https://github.com/ghc-proposals/ghc-proposals/pull/448, and click on the "rendered" link, namely https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0448-ty..., Github is now displaying the .rst source, not the rendered proposal.

This has started happening (consistently across other proposals) today.

Does anyone know why? How can I fix this?

Simon _______________________________________________ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs

Simon Peyton Jones

12 Jan 12 Jan

7:49 a.m.

But it used to work fine. What has changed? It seems like a pretty serious bug! On Thu, 11 Jan 2024 at 15:38, Rodrigo Mesquita wrote:

...

FWIW, I’m able to render/preview successfully other proposals, but 448-type-variable-scoping loads very slowly and ultimately fails to render for me too.

A comment in https://github.com/github/markup/issues/1016 (issue related to .rst not rendering) says that forcing a change in the file triggered a re-render successfully.

Rodrigo

On 11 Jan 2024, at 15:28, Simon Peyton Jones wrote:

If you go to https://github.com/ghc-proposals/ghc-proposals/pull/448, and click on the "rendered" link, namely https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0448-ty..., Github is now displaying the .rst source, not the rendered proposal.

This has started happening (consistently across other proposals) today.

Does anyone know why? How can I fix this?

Simon _______________________________________________ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs

Bryan Richter

19 Jan 19 Jan

7:16 a.m.

I'm guessing GitHub had to limit the size of .rst files that get rendered. I can imagine the kinds of annoying security or abuse issues they might be working around... The smallest file that doesn't work that I found by clicking around: https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0380-gh... (42KB) The largest that does still work: https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0581-na... (25KB) On Fri, 12 Jan 2024 at 14:49, Simon Peyton Jones < simon.peytonjones@gmail.com> wrote:

...

But it used to work fine. What has changed? It seems like a pretty serious bug!

On Thu, 11 Jan 2024 at 15:38, Rodrigo Mesquita < rodrigo.m.mesquita@gmail.com> wrote:

...
FWIW, I’m able to render/preview successfully other proposals, but 448-type-variable-scoping loads very slowly and ultimately fails to render for me too.

A comment in https://github.com/github/markup/issues/1016 (issue related to .rst not rendering) says that forcing a change in the file triggered a re-render successfully.

Rodrigo

On 11 Jan 2024, at 15:28, Simon Peyton Jones wrote:

If you go to https://github.com/ghc-proposals/ghc-proposals/pull/448, and click on the "rendered" link, namely https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0448-ty..., Github is now displaying the .rst source, not the rendered proposal.

This has started happening (consistently across other proposals) today.

Does anyone know why? How can I fix this?

Simon _______________________________________________ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs

_______________________________________________ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs

Simon Peyton Jones

10:07 a.m.

Maybe there's a config parameter that can be changed? It's extremely annoying, and a definite regression. Is there anyone at Github we can ask? Simon On Fri, 19 Jan 2024 at 12:16, Bryan Richter wrote:

...

I'm guessing GitHub had to limit the size of .rst files that get rendered. I can imagine the kinds of annoying security or abuse issues they might be working around...

The smallest file that doesn't work that I found by clicking around: https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0380-gh... (42KB)

The largest that does still work: https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0581-na... (25KB)

On Fri, 12 Jan 2024 at 14:49, Simon Peyton Jones < simon.peytonjones@gmail.com> wrote:

...
But it used to work fine. What has changed? It seems like a pretty serious bug!

On Thu, 11 Jan 2024 at 15:38, Rodrigo Mesquita < rodrigo.m.mesquita@gmail.com> wrote:

...
FWIW, I’m able to render/preview successfully other proposals, but 448-type-variable-scoping loads very slowly and ultimately fails to render for me too.

A comment in https://github.com/github/markup/issues/1016 (issue related to .rst not rendering) says that forcing a change in the file triggered a re-render successfully.

Rodrigo

On 11 Jan 2024, at 15:28, Simon Peyton Jones < simon.peytonjones@gmail.com> wrote:

If you go to https://github.com/ghc-proposals/ghc-proposals/pull/448, and click on the "rendered" link, namely https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0448-ty..., Github is now displaying the .rst source, not the rendered proposal.

This has started happening (consistently across other proposals) today.

Does anyone know why? How can I fix this?

Simon _______________________________________________ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs

_______________________________________________ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs

Ben Gamari

2:06 p.m.

Simon Peyton Jones writes:

...

Maybe there's a config parameter that can be changed? It's extremely annoying, and a definite regression. Is there anyone at Github we can ask?

If the problem is size then I rather doubt that there is anything that we can do about this beyond open a ticket and hope that someone notices. Indeed this sort of issue is why I have in the past been rather skeptical of relying on external services to host GHC's development. You can only expect to get the service that you pay for; in this case, that is nothing :) However, I am a bit curious as to why Rust's RFCs process has never run into this limitation. Perhaps it only affects RestructuredText? Cheers, - Ben

Adam Gundry

22 Jan 22 Jan

3:16 a.m.

I suspect this is a regression specific to reStructuredText rendering, but it's affecting a few different repositories [1] so hopefully Github will fix it at some point. In the meantime, for accepted proposals one can use https://ghc-proposals.readthedocs.io/en/latest/ which Joachim set up some time ago [2]. I'm not sure how frequently it is updated, as it doesn't have the GHC2024 proposal listed. And it sometimes tries to redirect to https://ghc-proposals.haskell.org which apparently doesn't exist. Joachim, do you know the status of this? (Of course, this doesn't help for rendering proposals that have not yet been accepted.) Cheers, Adam [1] https://github.com/orgs/community/discussions/86715 [2] https://mail.haskell.org/pipermail/ghc-devs/2019-May/017632.html On 19/01/2024 19:06, Ben Gamari wrote:

...

Simon Peyton Jones writes:

...
Maybe there's a config parameter that can be changed? It's extremely annoying, and a definite regression. Is there anyone at Github we can ask?

If the problem is size then I rather doubt that there is anything that we can do about this beyond open a ticket and hope that someone notices. Indeed this sort of issue is why I have in the past been rather skeptical of relying on external services to host GHC's development. You can only expect to get the service that you pay for; in this case, that is nothing :)

However, I am a bit curious as to why Rust's RFCs process has never run into this limitation. Perhaps it only affects RestructuredText?

Cheers,

- Ben

-- Adam Gundry, Haskell Consultant Well-Typed LLP, https://www.well-typed.com/ Registered in England & Wales, OC335890 27 Old Gloucester Street, London WC1N 3AX, England

Joachim Breitner

4:40 a.m.

Hi, Am Montag, dem 22.01.2024 um 08:16 +0000 schrieb Adam Gundry:

...

In the meantime, for accepted proposals one can use

https://ghc-proposals.readthedocs.io/en/latest/

which Joachim set up some time ago [2]. I'm not sure how frequently it is updated, as it doesn't have the GHC2024 proposal listed. And it sometimes tries to redirect to https://ghc-proposals.haskell.org which apparently doesn't exist. Joachim, do you know the status of this?

not sure if that subdomain ever existed, I reconfigured readthedocs to not refer to it. Also, they changed something recently, where I think they did not require a .readthedocs.yaml file and now they do. Let me add a simple one and see if it works again… Hmm, looks like the webhook integration also failed… updated. Ok, is updated again. But note that we never made that official, and more work would be needed to make that good (e.g. show the metadata, include the PR number in the main list), probably requiring some sphinx-python-hacking. And a decent domain. Cheers, Joachim -- Joachim Breitner mail@joachim-breitner.de http://www.joachim-breitner.de/

Joachim Breitner

4:56 a.m.

Hi, it’s not simply file size: we have (from wc -c proposals/*|sort -n) 43501 proposals/0380-ghc2021.rst 44051 proposals/0242-unsaturated-type-families.rst but #380 doesn't work but #242 works. I tried some rst linters, but nothing popped up. It could be output size or render time: It loading the file does take a while, as if github tries to render it and then gives up. That’s quite annoying, not sure what to do here now. Cheers, Joachim Am Freitag, dem 19.01.2024 um 14:16 +0200 schrieb Bryan Richter via ghc-devs:

...

I'm guessing GitHub had to limit the size of .rst files that get rendered. I can imagine the kinds of annoying security or abuse issues they might be working around...

The smallest file that doesn't work that I found by clicking around: https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0380-gh... (42KB)

The largest that does still work: https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0581-na... (25KB)

On Fri, 12 Jan 2024 at 14:49, Simon Peyton Jones wrote:

...
But it used to work fine. What has changed? It seems like a pretty serious bug!

On Thu, 11 Jan 2024 at 15:38, Rodrigo Mesquita wrote:

...
FWIW, I’m able to render/preview successfully other proposals, but 448-type-variable-scoping loads very slowly and ultimately fails to render for me too.

A comment in https://github.com/github/markup/issues/1016 (issue related to .rst not rendering) says that forcing a change in the file triggered a re-render successfully.

Rodrigo

...
On 11 Jan 2024, at 15:28, Simon Peyton Jones wrote:

If you go to https://github.com/ghc-proposals/ghc-proposals/pull/448, and click on the "rendered" link, namely https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0448-ty..., Github is now displaying the .rst source, not the rendered proposal.

This has started happening (consistently across other proposals) today.

Does anyone know why? How can I fix this?

Simon _______________________________________________ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs

_______________________________________________ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs

ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs

-- Joachim Breitner mail@joachim-breitner.de http://www.joachim-breitner.de/

Bryan Richter

8:22 a.m.

I cataloged all the proposals that fail to render. The others are ok. https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0111-li... https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0281-vi... https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0378-de... https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0380-gh... https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0448-ty... https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0583-ha... On Mon, 22 Jan 2024 at 11:56, Joachim Breitner wrote:

...

Hi,

it’s not simply file size: we have (from wc -c proposals/*|sort -n)

43501 proposals/0380-ghc2021.rst 44051 proposals/0242-unsaturated-type-families.rst

but #380 doesn't work but #242 works.

I tried some rst linters, but nothing popped up.

It could be output size or render time: It loading the file does take a while, as if github tries to render it and then gives up.

That’s quite annoying, not sure what to do here now.

Cheers, Joachim

...
I'm guessing GitHub had to limit the size of .rst files that get rendered. I can imagine the kinds of annoying security or abuse issues they might be working around...

The smallest file that doesn't work that I found by clicking around: https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0380-gh... (42KB)

The largest that does still work: https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0581-na... (25KB)

On Fri, 12 Jan 2024 at 14:49, Simon Peyton Jones < simon.peytonjones@gmail.com> wrote:

...
But it used to work fine. What has changed? It seems like a pretty serious bug!

On Thu, 11 Jan 2024 at 15:38, Rodrigo Mesquita < rodrigo.m.mesquita@gmail.com> wrote:

...
FWIW, I’m able to render/preview successfully other proposals, but 448-type-variable-scoping loads very slowly and ultimately fails to render for me too.

A comment in https://github.com/github/markup/issues/1016 (issue related to .rst not rendering) says that forcing a change in the file

Am Freitag, dem 19.01.2024 um 14:16 +0200 schrieb Bryan Richter via ghc-devs: triggered a re-render successfully.

...
...
...
Rodrigo

...
On 11 Jan 2024, at 15:28, Simon Peyton Jones <

simon.peytonjones@gmail.com> wrote:

...
...
If you go to

https://github.com/ghc-proposals/ghc-proposals/pull/448, and click on the "rendered" link, namely https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0448-ty..., Github is now displaying the .rst source, not the rendered proposal.

...
This has started happening (consistently across other proposals)

today.

...
Does anyone know why? How can I fix this?

Simon _______________________________________________ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs

_______________________________________________ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs

ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs

-- Joachim Breitner mail@joachim-breitner.de http://www.joachim-breitner.de/

_______________________________________________ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs

Bryan Richter

8:30 a.m.

This is strongly correlated with filesize. It's the list of the largest files in the repo, except #242 is missing. $ wc -c *.rst --total=never | sort -n | tail -n 7 43501 0380-ghc2021.rst 44051 0242-unsaturated-type-families.rst 55686 0448-type-variable-scoping.rst 56526 0583-hasfield-redesign.rst 56885 0281-visible-forall.rst 71114 0378-dependent-type-design.rst 92654 0111-linear-types.rst So yeah, perhaps it's actually limiting the size of the *output* or something. On Mon, 22 Jan 2024 at 15:22, Bryan Richter wrote:

...

I cataloged all the proposals that fail to render. The others are ok.

https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0111-li...

https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0281-vi...

https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0378-de...

https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0380-gh...

https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0448-ty...

https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0583-ha...

On Mon, 22 Jan 2024 at 11:56, Joachim Breitner wrote:

...
Hi,

it’s not simply file size: we have (from wc -c proposals/*|sort -n)

43501 proposals/0380-ghc2021.rst 44051 proposals/0242-unsaturated-type-families.rst

but #380 doesn't work but #242 works.

I tried some rst linters, but nothing popped up.

It could be output size or render time: It loading the file does take a while, as if github tries to render it and then gives up.

That’s quite annoying, not sure what to do here now.

Cheers, Joachim

...
I'm guessing GitHub had to limit the size of .rst files that get rendered. I can imagine the kinds of annoying security or abuse issues they might be working around...

The smallest file that doesn't work that I found by clicking around: https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0380-gh... (42KB)

The largest that does still work: https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0581-na... (25KB)

On Fri, 12 Jan 2024 at 14:49, Simon Peyton Jones < simon.peytonjones@gmail.com> wrote:

...
But it used to work fine. What has changed? It seems like a pretty serious bug!

On Thu, 11 Jan 2024 at 15:38, Rodrigo Mesquita < rodrigo.m.mesquita@gmail.com> wrote:

...
FWIW, I’m able to render/preview successfully other proposals, but 448-type-variable-scoping loads very slowly and ultimately fails to render for me too.

A comment in https://github.com/github/markup/issues/1016 (issue related to .rst not rendering) says that forcing a change in the file

...
...
...
Rodrigo

...
On 11 Jan 2024, at 15:28, Simon Peyton Jones <

simon.peytonjones@gmail.com> wrote:

...
...
If you go to

https://github.com/ghc-proposals/ghc-proposals/pull/448, and click on

Am Freitag, dem 19.01.2024 um 14:16 +0200 schrieb Bryan Richter via ghc-devs: triggered a re-render successfully. the "rendered" link, namely https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0448-ty..., Github is now displaying the .rst source, not the rendered proposal.

...
...
...
...
This has started happening (consistently across other proposals)

today.

...
Does anyone know why? How can I fix this?

Simon _______________________________________________ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs

_______________________________________________ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs

ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs

-- Joachim Breitner mail@joachim-breitner.de http://www.joachim-breitner.de/

_______________________________________________ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs

Simon Peyton Jones

11:15 a.m.

Is there any workaround at all? E.g. a web site where you can give it a URL for a RST file, and it'll typeset the RST for you? Or is it impossible to to typeset RST? Simon On Mon, 22 Jan 2024 at 13:22, Bryan Richter via ghc-devs < ghc-devs@haskell.org> wrote:

...

I cataloged all the proposals that fail to render. The others are ok.

https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0111-li...

https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0281-vi...

https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0378-de...

https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0380-gh...

https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0448-ty...

https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0583-ha...

On Mon, 22 Jan 2024 at 11:56, Joachim Breitner wrote:

...
Hi,

it’s not simply file size: we have (from wc -c proposals/*|sort -n)

43501 proposals/0380-ghc2021.rst 44051 proposals/0242-unsaturated-type-families.rst

but #380 doesn't work but #242 works.

I tried some rst linters, but nothing popped up.

It could be output size or render time: It loading the file does take a while, as if github tries to render it and then gives up.

That’s quite annoying, not sure what to do here now.

Cheers, Joachim

...
I'm guessing GitHub had to limit the size of .rst files that get rendered. I can imagine the kinds of annoying security or abuse issues they might be working around...

The smallest file that doesn't work that I found by clicking around: https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0380-gh... (42KB)

The largest that does still work: https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0581-na... (25KB)

On Fri, 12 Jan 2024 at 14:49, Simon Peyton Jones < simon.peytonjones@gmail.com> wrote:

...
But it used to work fine. What has changed? It seems like a pretty serious bug!

On Thu, 11 Jan 2024 at 15:38, Rodrigo Mesquita < rodrigo.m.mesquita@gmail.com> wrote:

...
FWIW, I’m able to render/preview successfully other proposals, but 448-type-variable-scoping loads very slowly and ultimately fails to render for me too.

A comment in https://github.com/github/markup/issues/1016 (issue related to .rst not rendering) says that forcing a change in the file

...
...
...
Rodrigo

...
On 11 Jan 2024, at 15:28, Simon Peyton Jones <

simon.peytonjones@gmail.com> wrote:

...
...
If you go to

https://github.com/ghc-proposals/ghc-proposals/pull/448, and click on

Am Freitag, dem 19.01.2024 um 14:16 +0200 schrieb Bryan Richter via ghc-devs: triggered a re-render successfully. the "rendered" link, namely https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0448-ty..., Github is now displaying the .rst source, not the rendered proposal.

...
...
...
...
This has started happening (consistently across other proposals)

today.

...
Does anyone know why? How can I fix this?

Simon _______________________________________________ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs

_______________________________________________ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs

ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs

-- Joachim Breitner mail@joachim-breitner.de http://www.joachim-breitner.de/

_______________________________________________ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs

_______________________________________________ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs

Adam Gundry

3:48 p.m.

On 22/01/2024 16:15, Simon Peyton Jones wrote:

...

Is there any workaround at all? E.g. a web site where you can give it a URL for a RST file, and it'll typeset the RST for you? Or is it impossible to to typeset RST?

I tend to use pandoc to preview RST files locally. There are also a few websites that allow you to enter RST content directly and see a live preview (e.g. https://snippets.documatt.com/). They may not exactly line up with how GitHub renders things, of course, and I don't know of any that let you feed in a URL to an RST file hosted elsewhere. Adam

...

On Mon, 22 Jan 2024 at 13:22, Bryan Richter via ghc-devs mailto:ghc-devs@haskell.org> wrote:

I cataloged all the proposals that fail to render. The others are ok.

https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0111-li... https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0111-li... https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0281-vi... https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0281-vi... https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0378-de... https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0378-de... https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0380-gh... https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0380-gh... https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0448-ty... https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0448-ty... https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0583-ha... https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0583-ha...

On Mon, 22 Jan 2024 at 11:56, Joachim Breitner mailto:mail@joachim-breitner.de> wrote:

Hi,

it’s not simply file size: we have (from wc -c proposals/*|sort -n)

43501 proposals/0380-ghc2021.rst 44051 proposals/0242-unsaturated-type-families.rst

but #380 doesn't work but #242 works.

I tried some rst linters, but nothing popped up.

It could be output size or render time: It loading the file does take a while, as if github tries to render it and then gives up.

That’s quite annoying, not sure what to do here now.

Cheers, Joachim

Am Freitag, dem 19.01.2024 um 14:16 +0200 schrieb Bryan Richter via ghc-devs: > I'm guessing GitHub had to limit the size of .rst files that get rendered. I can imagine the kinds of annoying security or abuse issues they might be working around... > > The smallest file that doesn't work that I found by clicking around: https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0380-gh... https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0380-gh... (42KB) > > The largest that does still work: https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0581-na... https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0581-na... (25KB) > > On Fri, 12 Jan 2024 at 14:49, Simon Peyton Jones mailto:simon.peytonjones@gmail.com> wrote: > > But it used to work fine. What has changed? It seems like a pretty serious bug! > > > > On Thu, 11 Jan 2024 at 15:38, Rodrigo Mesquita mailto:rodrigo.m.mesquita@gmail.com> wrote: > > > FWIW, I’m able to render/preview successfully other proposals, but 448-type-variable-scoping loads very slowly and ultimately fails to render for me too. > > > > > > A comment in https://github.com/github/markup/issues/1016 https://github.com/github/markup/issues/1016 (issue related to .rst not rendering) says that forcing a change in the file triggered a re-render successfully. > > > > > > Rodrigo > > > > > > > > > > On 11 Jan 2024, at 15:28, Simon Peyton Jones mailto:simon.peytonjones@gmail.com> wrote: > > > > > > > > If you go to https://github.com/ghc-proposals/ghc-proposals/pull/448 https://github.com/ghc-proposals/ghc-proposals/pull/448, and click on the "rendered" link, namely https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0448-ty... https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0448-ty..., Github is now displaying the .rst source, not the rendered proposal. > > > > > > > > This has started happening (consistently across other proposals) today. > > > > > > > > Does anyone know why? How can I fix this? > > > > > > > > Simon

-- Adam Gundry, Haskell Consultant Well-Typed LLP, https://www.well-typed.com/ Registered in England & Wales, OC335890 27 Old Gloucester Street, London WC1N 3AX, England

Ben Gamari

8:30 p.m.

Adam Gundry writes:

...

On 22/01/2024 16:15, Simon Peyton Jones wrote:

...
Is there any workaround at all? E.g. a web site where you can give it a URL for a RST file, and it'll typeset the RST for you? Or is it impossible to to typeset RST?

I tend to use pandoc to preview RST files locally. There are also a few websites that allow you to enter RST content directly and see a live preview (e.g. https://snippets.documatt.com/). They may not exactly line up with how GitHub renders things, of course, and I don't know of any that let you feed in a URL to an RST file hosted elsewhere.

As noted during the SWG meeting today, I suspect that the easiest way forward here may be to just render PRs via CI. However, I don't know how easy it is to preserve artifacts such that they are viewable using GitHub Actions. Jose will have a quick look at this. Cheers, - Ben

Bryan Richter

23 Jan 23 Jan

1:12 a.m.

One really hacky solution that might work would just be to split a proposal across two files. On Tue, 23 Jan 2024 at 03:31, Ben Gamari wrote:

...

Adam Gundry writes:

...
On 22/01/2024 16:15, Simon Peyton Jones wrote:

...
Is there any workaround at all? E.g. a web site where you can give it a URL for a RST file, and it'll typeset the RST for you? Or is it impossible to to typeset RST?

I tend to use pandoc to preview RST files locally. There are also a few websites that allow you to enter RST content directly and see a live preview (e.g. https://snippets.documatt.com/). They may not exactly line up with how GitHub renders things, of course, and I don't know of any that let you feed in a URL to an RST file hosted elsewhere.

As noted during the SWG meeting today, I suspect that the easiest way forward here may be to just render PRs via CI. However, I don't know how easy it is to preserve artifacts such that they are viewable using GitHub Actions. Jose will have a quick look at this.

Cheers,

- Ben

_______________________________________________ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs

Tom Ellis

3:36 a.m.

On Mon, Jan 22, 2024 at 08:30:47PM -0500, Ben Gamari wrote:

...

...
On 22/01/2024 16:15, Simon Peyton Jones wrote:

...
Is there any workaround at all? E.g. a web site where you can give it a URL for a RST file, and it'll typeset the RST for you? Or is it impossible to to typeset RST?

I tend to use pandoc to preview RST files locally. There are also a few websites that allow you to enter RST content directly and see a live preview (e.g. https://snippets.documatt.com/). They may not exactly line up with how GitHub renders things, of course, and I don't know of any that let you feed in a URL to an RST file hosted elsewhere.

As noted during the SWG meeting today, I suspect that the easiest way forward here may be to just render PRs via CI. However, I don't know how easy it is to preserve artifacts such that they are viewable using GitHub Actions. Jose will have a quick look at this.

Perhaps the artifacts could be deployed as Github Pages. But in the first instance I like Bryan's suggestion of just splitting the documents into small-enough pieces.

Michael Peyton Jones

5:16 a.m.

If you got the readthedocs site working, that does provide per-PR previews of the built documentation relatively easily. M On Tue, Jan 23, 2024 at 8:37 AM Tom Ellis < tom-lists-haskell-cafe-2023@jaguarpaw.co.uk> wrote:

...

...
...
On 22/01/2024 16:15, Simon Peyton Jones wrote:

...
Is there any workaround at all? E.g. a web site where you can give it a URL for a RST file, and it'll typeset the RST for you? Or is it impossible to to typeset RST?

I tend to use pandoc to preview RST files locally. There are also a few websites that allow you to enter RST content directly and see a live preview (e.g. https://snippets.documatt.com/). They may not exactly

On Mon, Jan 22, 2024 at 08:30:47PM -0500, Ben Gamari wrote: line

...
...
up with how GitHub renders things, of course, and I don't know of any that let you feed in a URL to an RST file hosted elsewhere.

As noted during the SWG meeting today, I suspect that the easiest way forward here may be to just render PRs via CI. However, I don't know how easy it is to preserve artifacts such that they are viewable using GitHub Actions. Jose will have a quick look at this.

Perhaps the artifacts could be deployed as Github Pages. But in the first instance I like Bryan's suggestion of just splitting the documents into small-enough pieces. _______________________________________________ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs

Joachim Breitner

4:36 p.m.

Hi, yes, I enabled that PR preview feature yesterday, but it seems that I also have to grand permissions to their OAuth App to the ghc-proposals github org … done. Hmm, now every PR that isn’t based on a recent master is red, because it lacks the configuration. Also not nice. Maybe I should disable this again? At least until the readthedocs setup and output is satisfying (numbering etc. – any volunteers?) Cheers, Joachim Am Dienstag, dem 23.01.2024 um 10:16 +0000 schrieb Michael Peyton Jones:

...

...
If you got the readthedocs site working, that does provide per-PR previews of the built documentation relatively easily.

M

On Tue, Jan 23, 2024 at 8:37 AM Tom Ellis wrote:

...
...
On Mon, Jan 22, 2024 at 08:30:47PM -0500, Ben Gamari wrote:

...
...
> > On 22/01/2024 16:15, Simon Peyton Jones wrote: > > > > Is there any workaround at all? E.g. a web site where you can give it > > > > a URL for a RST file, and it'll typeset the RST for you? Or is it > > > > impossible to to typeset RST? > > > > I tend to use pandoc to preview RST files locally. There are also a few > > websites that allow you to enter RST content directly and see a live > > preview (e.g. https://snippets.documatt.com/). They may not exactly line > > up with how GitHub renders things, of course, and I don't know of any > > that let you feed in a URL to an RST file hosted elsewhere. > > As noted during the SWG meeting today, I suspect that the easiest way forward here may be to just render PRs via CI. However, I don't know how easy it is to preserve artifacts such that they are viewable using GitHub Actions. Jose will have a quick look at this.

Perhaps the artifacts could be deployed as Github Pages. But in the first instance I like Bryan's suggestion of just splitting the documents into small-enough pieces. _______________________________________________ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs

ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs

-- Joachim Breitner mail@joachim-breitner.de http://www.joachim-breitner.de/

Ganesh Sittampalam

22 Mar 22 Mar

8:53 p.m.

I believe the problem is fixed now - at the very least, the one in Simon's original report is now working. On 11/01/2024 15:28, Simon Peyton Jones wrote:

...

If you go to https://github.com/ghc-proposals/ghc-proposals/pull/448, and click on the "rendered" link, namely https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0448-ty..., Github is now displaying the .rst source, not the rendered proposal.

This has started happening (consistently across other proposals) today.

Does anyone know why? How can I fix this?

Simon

_______________________________________________ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs

423

Age (days ago)

495

Last active (days ago)

List overview

Download

18 comments

9 participants

participants (9)

Adam Gundry
Ben Gamari
Bryan Richter
Ganesh Sittampalam
Joachim Breitner
Michael Peyton Jones
Rodrigo Mesquita
Simon Peyton Jones
Tom Ellis

GHC proposals not rendering

tags

participants (9)