| Age | Commit message (Collapse) | Author |
|---|
|
(Partially reverts commit e0bb4e0.)
Actually fix generate_name() COUNT query this time. It still had an
imaginary table "pomf" from e0bb4e0. Change it to "files" like in the
schema.
For code consistency, :filename is reverted back to :name as it was
before commit e0bb4e0. That commit is bogus and doesn't fix the issue
due to skipping the code review process.
The problem with this COUNT query in first place was that it checked
imaginary columns (name) from an imaginary table (pomf) in the database.
This led to the "same filename" check not working at all, and the other
code only relying on checking the same hash (currently SHA-1) for
returning the filename before generate_name().
By luck, generated filenames could collide and another file of the same
generated filename would be returned, discarding the new upload.
I discovered this bug about a month ago myself, but back then there was
no reliable way to reproduce it and no good description of the bug.
This bug was introduced in commit 5b77a4e.
Reported-by: Vhum Vhum <vhumvhum1@gmail.com>
Suggested-by: Luminarys <luminarys@fuwa.se>
OK ewhal@, luminarys@
|
|
database
|
|
Hidden <input> values did not work as I expected for POST requests. In
GET requests, the hidden values are added as query parameters to the
URL. For POST requests, the hidden inputs are simply more form-data to
upload in the POST request (another "file").
Because of this, the HTML output parameter is never passed to upload.php
and so it returns the default response instead (JSON). The goal is to
use the HTML response.
This reverts commit d6dc103c7229cb0e2e177ce0dfc0cfb528025ca9, where the
bug was introduced.
OK ewhal@
|
|
Bump grunt-contrib-copy ^0.8.1 -> ^0.8.2. While this is not entirely
necessary, do it anyway to keep our package.json up to date with latest
releases.
Bump grunt-contrib-uglify ^0.9.2 -> ^0.10.0. Updates UglifyJS dependency
from ^2.4.24 -> ^2.5. May result in better, smaller or less buggy
minified JavaScript, although this has not been tested.
If everything goes well, this commit doesn't change anything. Carets
will automatically update to latest major version. This commit just
means we are showing to drop support for some older versions, despite
package.json really instructing to download a newer release.
OK ewhal@
|
|
Clean up CSS. Removes an off-screen content hack and removes
unnecessary, dead style rules.
POCKY~ rules have never seen light in the source code, not in core nor
moe panel.
CSS border-box vendor prefixes are long since unnecessary and border-box
is supported natively by browsers.
OK ewhal@
|
|
Add .mailmap, which is a git-shortlog(1) feature. Maps author names with
different names/email addresses under one name.
OK ewhal@
|
|
The message and the results for why the law enforcement must be
contacted may be unclear to some. Add "they may contact us if
appropriate" to clarify the message and uphold the legal spirit, for the
lack of better words.
|
|
As seen on the ShareX project's home page, the correct spelling of the
name is with an uppercase `X`. On the tools page, it is incorrectly an
lowercase `x` (Sharex). This corrects it.
|
|
The pantsucat/pomfsecure repository has been dead for a very long time.
It is giving a "404 Not Found" error. Fix the download link to a working
Git repository.
|
|
- Revert project name change made in commit 0ab45bf3 back from Pantsu
back to Pomf.
- Change package version to 2.0.0-dev (one time early version change
exception).
- Change description's initial character to camelcase.
|
|
ewhal@ didn't understand how changing the maximum file size works, so he
edited the JavaScript code directly to set the default to his desired
120MiB.
For all clones, this may not be viable and it's not a sane default in my
opinion. This doesn't change the fact that the maximum file size remains
to be what it's configured in the configuration (Gruntfile.js).
This bug was introduced in commit ec0154b, and potentially others.
Vague OK from ewhal@ many weeks ago.
|
|
Problems with placing #upload-input off-screen is that keyboard
navigation (tab) will gain focus of the upload buttons twice.
Hiding the other upload button (the <input> element) off-screen makes no
sense because no explanation is given. Hide it with `display: none;`.
Considering screen reader users, the <button> element (with a future
<label> element) should hopefully convey enough information about it's
purpose. `tabindex="-1"` would be an alternative, but would require
placing content off-screen.
|
|
border-box has been supported natively in web browsers since Chromium
10, Firefox 29.0, Internet Explorer 8, Opera 7, Safari (WebKit) 5.1. IE
never also had a vendor prefix.
As a guideline, as of today Firefox ESR (extended support release) is on
version 38.3.0, so native support has been available for a long while.
These vendor prefixes are no longer needed. This commit removes them.
For future improvement, matching this rule against wildcard `*` should
be re-considered and moved to body or some other rule to avoid expensive
CSS calculations.
|
|
Pomf core nor moe panel have never used these POCKY~ style rules. They
have no reason to exist. Remove them for cleanup.
Reported-by: Luminarys <luminarys@fuwa.se>
|
|
There are commented style rules in the CSS that do not appear to have
been used in a long time, and have no reason to exist. Remove them for
cleanup.
|
|
Update the FAQ to be semantically more correct, easier to read and
pretty.
A lot of here applies to Pantsu.cat specifically. For future
improvement, a generic FAQ may be created and this FAQ split into an
extension repository (pomf-pages-pantsucat or similar).
OK ewhal@
|
|
Although alchimist has not yet merged any commits to master as of this
commit, it is anticipated to happen eventually after he's done working
with his branches. Regardless, adding this makes sure git-shortlog(1)
will display the up-to-date email address.
|
|
Eliot has also contributed three commits with his name using
contact@pantsu.cat. The match on pantsucat name is too specific.
This groups those three commits to Eliot Whalan <ewahl@pantsu.cat>.
|
|
From git-shortlog(1) man page:
> The .mailmap feature is used to coalesce together commits by the same
> person in the shortlog, where their name and/or email address was
> spelled differently.
|
|
A "valid notice" for copyright takedown is questionable and doesn't tell
much of a clue what needs to be in a takedown notice. Change it to
specify Copyright Act 1968 (Australia), which is the current copyright
legislation in Australia.
For future improvement, a copyright takedown page could be created with
more detailed instructions what needs to be included in a copyright
takedown notice.
|
|
nav.swig includes were inconsistently at either bottom of the page or
inside the jumbotron. For FAQ, the navigation was also nested
incorrectly inside the <article> element, which means the navigation
incorrectly assumed to be in-article navigation (table of contents).
The intention is to have these nav.swig (site-wide navigation) includes
outside of <article> scope. Future improvement is desired to have all
pages like other.swig and tools.swig to have <article> elements.
|
|
The <article> element is an element introduced in HTML5 for semantic
sectioning.
faq.swig is missing this semantic element, so this commit adds it.
It doesn't seem appropriate to use <article> in layout.swig, because
it's used for every page, also ones that are not articles.
Reference: http://www.w3.org/TR/html5/sections.html#the-article-element
|
|
The previous paragraph is misleading. On takedown notice, the intention
is not to disable access to files the complainer has uploaded. This
fixes the paragraph to say the intention is to disable access to
infringing content.
|
|
Me and antonizoon have decided to create a new central channel for
Pomf.se clone discussion. This leaves #/g/hostingtalk dedicated to its
intended purpose, general server hosting discussion.
The channel name #pomfret comes from ArchiveTeam's EFnet channel of the
same name, which was used to rescue Pomf.se's files. The channel on
EFnet is without OPs these days. antonizoon also suggested lots of
imageboard/chan channels are also on Rizon, so Rizon was chosen for the
channel as a good fit. Rizon also houses the #bibanon channel, which is
the official channel for Bibliotheca Anonoma.
OK ewhal@
|
|
The banner in FAQ seems out of place and doesn't really belong to the
page. This commit removes the banners from faq.swig. The banners are
still kept on the index page.
|
|
The previous FAQ is difficult to read and doesn't actually answer to
common questions. It also lacks semantics.
Because the previous FAQ had reached a point of being next to useless
and unfixable with minor edits, it had to be rewritten.
This commit addresses the most commonly asked questions. It uses HTML5
headings for addresssing each question for proper document outline. For
future improvement, more may be added as necessary.
|
|
FAQ itself as a word as unsemantic to screen readers and not accessible.
Adding an <abbr> element improves accessibility and makes screen readers
read the title correctly.
|
|
Peter didn't actually write this piece of Swig template, and there's an
insignificant amount of templating code that could have been written by
him. We can remove him from this template's authors.
|
|
The link to real world example at Pantsu.cat is not semantically a
feature, but its own demonstration. Adding a subheader for the example
increases semantics further, and makes it easier to find.
|
|
The README.md lacks a features section, which is worthful for those
searching for a suitable file sharing platform to host.
This change lists the features and seperates "Requirements" to its own
section.
Worthy read: "How to Write a Readme Worth Reading" by Diana Thayer.
https://orchestrate.io/blog/2014/07/16/how-to-write-a-readme-worth-reading/
|
|
This bug was introduced in previous commit 566611c.
|
|
This is not the right way to do i18n. Banners are difficult and buggy
with translations, the folder structure is wrong and the result is
duplicated markup.
For the sake of maintainability, I am liberally removing this from core.
Future i18n attempts should use real gettext implementations and
possibly seperate i18n files from core as an extension.
Because Cupkle contributed the Catalan translation, this also removes
him/her from contributors.
|
|
WCAG 2.0 section 1.4.3 sets certain criteria for foreground/background
color contrast ratio. The current contrast ratio is too low at 4.00,
where the desired contrast ratio is at least 4.50.
This commit changes the contrast ratio to the minimum required to
fulfill the WCAG 2.0 section 1.4.3 criteria.
Previously this was partially addressed in commit 6915a50, which neku
forgot to attribute to my name.
Reference:
http://www.w3.org/TR/UNDERSTANDING-WCAG20/visual-audio-contrast-contrast.html
|
|
This does nothing as far as I see it, no dist/img/ is created. And hey,
it works without it. grunt-contrib-copy can create directories for us.
|
|
Grunt would not copy `img/` to `dist/img/`, because the task of copying
images was previously on the now-removed imagemin task.
This bug was introduced in commit 90e6b7b.
|
|
The benefits of HTML minification are too small and outweighted by the
benefits of not having to deal with yet another dependency in Pomf's
code. `Vary: Accept-Encoding` on the server is the recommended way to
approach this, also for clarity and debugging without browser consoles.
We don't need this.
|
|
Ever since liberating images in c76f70c, there's no really a need to
minify images. Furthermore, the dependency update in 6ef710a changed PNG
minification to be lossless, so most of the benefits were lost (because
our images are already compressed in the repository).
We don't need this anymore.
|
|
JSHint is "deprecated" in favor of ESLint. Personally I have not used
JSHint ever since switching to ESLint.
Vague OK from ewhal@
|
|
To understand and document the API, responses should be correct,
accurate and human readable. Server-side compression (gzip) will
compensate for the extra whitespace.
OK ewhal@
|
|
Adds "Contributing", and "Credits" sections. Rewrites "Configuring".
Replaces "Contact" with "Getting help".
OK alchimist@
|
|
Thanks to alchimist@ for doing editorial review.
|
|
The previous example can create a feeling of worry in the mind of the
reader. This will aim to fix that.
|
|
We use IRC for code review, so naturally merge requests also happen
there.
|
|
|
|
Worthy read: "Top ten reasons why I won’t use your open source project"
by Adam Stacoviak.
https://changelog.com/top-ten-reasons-why-i-wont-use-your-open-source-project/
|
|
It seems neku doesn't have a working HTTPS website anymore. It certainly
used to.
|
|
|
|
This also adds a mention of the IRC channel.
Worthy read: "Readme Driven Development" by Tom Preston-Werner:
http://tom.preston-werner.com/2010/08/23/readme-driven-development.html
|
|
Fixes a dead link.
|
|
Before this merge, ESLint reported about 114 problems (76 errors, 38
warnings). This merge reduces that to about 67 problems (29 errors, 38
warnings) using Airbnb ES5 config.
OK ewhal@
|
