Adds side-by-side diff for images

[saitho]

This adds a basic image diff and prints the width of the images (before and after). This is part of what was requested in #6255 .

Screenshot:

The side-by-side view for images will be active regardless whether split view is active or not.

[codecov-io]

Codecov Report

Merging #6784 into master will decrease coverage by 0.04%.
The diff coverage is 8.77%.

@@            Coverage Diff             @@
##           master    #6784      +/-   ##
==========================================
- Coverage   41.64%   41.59%   -0.05%     
==========================================
  Files         483      483              
  Lines       64631    64741     +110     
==========================================
+ Hits        26916    26932      +16     
- Misses      34274    34360      +86     
- Partials     3441     3449       +8

Impacted Files	Coverage Δ
modules/git/commit.go	51.89% <0%> (-5.31%)	⬇️
routers/repo/commit.go	27.03% <0%> (-2%)	⬇️
routers/repo/pull.go	30.74% <0%> (-0.88%)	⬇️
routers/repo/compare.go	48.74% <21.27%> (-5.11%)	⬇️
routers/repo/view.go	43.54% <0%> (+1.01%)	⬆️
modules/log/event.go	65.64% <0%> (+1.02%)	⬆️
models/unit.go	67.56% <0%> (+5.4%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update a5f87fe...69083b1. Read the comment docs.

[lafriks]

Probably better would be to have Width: 1156, Height: 412 instead of W: 1156, H: 412

Also what will happen with very large images?

[saitho]

The images are embedded how they were uploaded in the repository, i.e. in their full dimensions.
The width is limited by width: 100% in .diff-file-box.file-content img which is already present.

[saitho]

Note that I changed "headTarget" to "baseTarget", as the "before" paths should point to the forked repository.

[silverwind]

If it's not too hard, I'd request adding a size below each image. Would be useful to compare size optimizations:

[saitho]

If it's not too hard, I'd request adding a size below each image. Would be useful to compare size optimizations

I'll look into that. :) I assume that the filesystem library is able to do that.

[saitho]

There we go. :)

[saitho]

Okay. I'll do that on the weekend, as the PR isn't relevant for 1.9.0. ;) Jonas Franz <[email protected]> schrieb am Di., 9. Juli 2019, 22:21:

…

***@***.**** requested changes on this pull request. ------------------------------ In routers/repo/pull.go <#6784 (comment)>: > ctx.Data["RawPath"] = setting.AppSubURL + "/" + path.Join(headTarget, "raw", "commit", endCommitID) + ctx.Data["BeforeSourcePath"] = setting.AppSubURL + "/" + path.Join(baseTarget, "src", "commit", startCommitID) Missing check? What happens if there is no startCommitID because it's the first commit? (see my last comment) ------------------------------ In routers/repo/pull.go <#6784 (comment)>: > ctx.Data["RawPath"] = setting.AppSubURL + "/" + path.Join(headTarget, "raw", "commit", endCommitID) + ctx.Data["BeforeSourcePath"] = setting.AppSubURL + "/" + path.Join(baseTarget, "src", "commit", startCommitID) + ctx.Data["BeforeRawPath"] = setting.AppSubURL + "/" + path.Join(baseTarget, "raw", "commit", startCommitID) same — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#6784?email_source=notifications&email_token=ABXUS7SZ677NYBA2C4CCDA3P6TXNVA5CNFSM4HI5BMVKYY3PNVWWK3TUL52HS4DFWFIHK3DMKJSXC5LFON2FEZLWNFSXPKTDN5WW2ZLOORPWSZGOB55UPXA#pullrequestreview-259737564>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ABXUS7SGVLDQFQH7JBKNVR3P6TXNVANCNFSM4HI5BMVA> .

[saitho]

Hey there. Sorry for the long delay.
Wasn't really motivated to work on Gitea - now I am. :)

2. When I create a new empty repository and commit an image in the first commit, there will  be a corrupted old image. (see screenshot and code)
   ![Screenshot_2019-07-09 geissler](https://user-images.githubusercontent.com/5757182/60919844-063e0500-a297-11e9-9304-4f0343286b00.png)

This happens when the image is newly added to the repository (i.e. is not in the index of the base commit). Fixed by only displaying the image if it exists in the base commit index.

[saitho]

Ready for review @jonasfranz

[guillep2k]

Could this PR have latest master merged in and CI restarted? Looks like CI problems are not from this PR.

[lunny]

Since @jonasfranz didn't response this one for a month and there are two approvals on this, I think we can merge this one when CI PASS.

As per @lunny comment

[guillep2k]

@lunny @lafriks It looks like the tests I've added to the UTF functions are quite problematic. They fail only from time to time. I guess it's a LOCALE setting. Can a specific LOCALE be forced in the Makefile for the tests? Something like LOCALE=en_US.UTF-8 or any other, as long as it is always the same.

[guillep2k]

Otherwise, I'd remove most of the UTF8 tests. I don't think they're that much important.

[lunny]

make L-G-T-M work