Re: Upload leads to blank page
| From: | Derick Rethans | Date: | Tue, 05 Mar 2024 09:43:54 +0000 |
| Subject: | Re: Upload leads to blank page | ||
| References: | 1 2 3 | Groups: | php.pecl.dev |
| Request: | Send a blank email to pecl-dev+get-17572@lists.php.net to get a copy of this message | ||
On Tue, 5 Mar 2024, Tim Starling wrote:
> On 29/2/24 08:12, Derick Rethans wrote:
> > On 28 February 2024 01:48:15 GMT, Tim Starling <tstarling@wikimedia.org>
> > wrote:
> > > When I uploaded the new release of excimer, submitting the verify
> > > page responded with a blank page.
> > >
> > > I tried opening developer tools and reloading the page, but it
> > > reported that the release already existed.
> > >
> > > There was probably a fatal error late in the release process.
> > >
> > > I thought you would like to know.
> > >
> > > -- Tim Starling
> > Thanks.
> >
> > Can you try to make sure not to have an UTF-8, but just ASCII in
> > your package.xml file? Can you send me the tar ball to try? You can
> > also try deleting the uploaded release, and then retry with just
> > that ASCII in package.xml.
>
> Was my followup message not delivered? I wrote:
>
> > This was probably because the package.xml had an error in it which
> > was missed by "pear package-validate" but caused a fatal error in
> > Dependency2::validateOsDependency() when installing the package.
> > Remi reported the issue to us and proposed a fix.
> >
> > https://github.com/wikimedia/mediawiki-php-excimer/pull/7
It did come through, and I had already added an item to my todo list to
check it out. I just haven't had the time yet to look at it :-).
> Remi's fix was merged and released in excimer 1.2.1 which was uploaded
> successfully.
>
> Really it's a PEAR validator bug. Validation should have detected the
> issue before it became a fatal error.
Yes, but the chances of fixing *PEAR* are probably pretty low now.
> The broken package, 1.2.0, was uploaded and you can find it at
> https://pecl.php.net/get/excimer-1.2.0.tgz , it's just
> that there was
> an error after upload, late in the request.
This is what I want to have a look at still. The fatal error should have
aborted the upload, and not leave things in a broken state.
cheers,
Derick