`sdl2_image`: fix downloading #3135

T-Dynamos · 2025-03-26T04:32:13Z

[DEBUG]:        Cloning into 'external/libpng'...
[DEBUG]:        remote: Enumerating objects: 109787, done.
remote: Counting objects: 100% (11/11), done.
remote: Compressing objects: 100% (6/6), done.
error: RPC failed; curl 92 HTTP/2 stream 5 was not closed cleanly: CANCEL (err 8)
[DEBUG]:        error: 3705 bytes of body are still expected
[DEBUG]:        fetch-pack: unexpected disconnect while reading sideband packet
[DEBUG]:        fatal: early EOF
[DEBUG]:        fatal: fetch-pack: invalid index-pack output

Ref: #3115 (comment)

misl6 · 2025-04-08T19:26:09Z

Tried a build (without this PR) a few hours ago and worked flawlessly, is this patch still required on your side?

T-Dynamos · 2025-04-09T05:00:15Z

This happens on slow internet connections only. Also it prevents download of commit history for all submodules.

AndiEcker · 2025-04-09T12:17:14Z

Confirming that this bug is happening to me sometimes, when I use buildozer/p4a in my home office which is using a slow ethernet-via-powerline connection. Am Mi., 9. Apr. 2025 um 06:00 Uhr schrieb Ansh Dadwal < ***@***.***>:

…

This happens on slow internet connections only. Also it prevents download of commit history for all submodules. — Reply to this email directly, view it on GitHub <#3135 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AA4U63Q3MWCQ4TLO2UB4V2D2YSSP3AVCNFSM6AAAAABZZIVD4CVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDOOBYGI4DIMZVGY> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***> *T-Dynamos* left a comment (kivy/python-for-android#3135) <#3135 (comment)> This happens on slow internet connections only. Also it prevents download of commit history for all submodules. — Reply to this email directly, view it on GitHub <#3135 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AA4U63Q3MWCQ4TLO2UB4V2D2YSSP3AVCNFSM6AAAAABZZIVD4CVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDOOBYGI4DIMZVGY> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

The end of a `with` block already does that.

T-Dynamos · 2025-05-12T16:57:36Z

@clayote Thanks for taking over this.

Please do confirm if same fix is needed for sdl3_image.

clayote · 2025-05-12T16:59:33Z

I don't know a lot about sdl3 or even 2, really. The PR looks good, and it only affects the downloading stage, so I figured I'd tidy it a bit.

clayote · 2025-05-12T17:03:46Z

@AndiEcker I don't have a slow connection to test on, could you confirm that this PR fixes your issue?

AndiEcker · 2025-05-13T10:48:12Z

@clayote: sorry I am not sure how to test this - I tried to put the following lines into my buildozer.spec: p4a.branch = T-Dynamos:sdl2_image_fix or p4a.branch = sdl2_image_fix but in both cases I get only an branch-not-exists-error (because your pull request branch is only available in your fork and not in the P4A repo) Am Mo., 12. Mai 2025 um 18:04 Uhr schrieb clayote ***@***.***

…

: *clayote* left a comment (kivy/python-for-android#3135) <#3135 (comment)> @AndiEcker <https://github.com/AndiEcker> I don't have a slow connection to test on, could you confirm that this PR fixes your issue? — Reply to this email directly, view it on GitHub <#3135 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AA4U63QHVZV5FHT3MDL52VD26DIARAVCNFSM6AAAAABZZIVD4CVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDQNZTGMZDIOJQGA> . You are receiving this because you were mentioned.Message ID: ***@***.***>

T-Dynamos · 2025-05-13T10:52:40Z

@AndiEcker

p4a.fork = T-Dynamos
p4a.branch = sdl2_image_fix

AndiEcker · 2025-05-13T11:11:19Z

Thanks for your help!! silly me - totally forgot the p4a.fork option I just started the first test build run (needs around 40 minutes). I will have to do maybe 4 to 6 test build runs before I come back here, because the fixed bug did not occur on each build run. Am Di., 13. Mai 2025 um 11:53 Uhr schrieb Ansh Dadwal < ***@***.***>:

…

*T-Dynamos* left a comment (kivy/python-for-android#3135) <#3135 (comment)> @AndiEcker <https://github.com/AndiEcker> p4a.fork = T-Dynamos — Reply to this email directly, view it on GitHub <#3135 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AA4U63RK6IG5OZI4JB5CD6326HFI5AVCNFSM6AAAAABZZIVD4CVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDQNZWGAYDQMBVGE> . You are receiving this because you were mentioned.Message ID: ***@***.***>

AndiEcker · 2025-05-14T15:11:49Z

Yesterday and today I did 7 test build runs and tried hard (adding parallel downloads to overcharge my slow power line ethernet connection, so that the buildozer build time got doubled), but could not reproduce any (early EOF) download errors! Before your fix I got at least 1 build fail out of 3..4 runs. So it looks like your PR fixed this bug.!!! ¡¡¡ Thank you very much for your great work !!! Am Di., 13. Mai 2025 um 12:10 Uhr schrieb Andreas Ecker ***@***.***>:

…

Thanks for your help!! silly me - totally forgot the p4a.fork option I just started the first test build run (needs around 40 minutes). I will have to do maybe 4 to 6 test build runs before I come back here, because the fixed bug did not occur on each build run. Am Di., 13. Mai 2025 um 11:53 Uhr schrieb Ansh Dadwal < ***@***.***>: > *T-Dynamos* left a comment (kivy/python-for-android#3135) > <#3135 (comment)> > > @AndiEcker <https://github.com/AndiEcker> > > p4a.fork = T-Dynamos > > — > Reply to this email directly, view it on GitHub > <#3135 (comment)>, > or unsubscribe > <https://github.com/notifications/unsubscribe-auth/AA4U63RK6IG5OZI4JB5CD6326HFI5AVCNFSM6AAAAABZZIVD4CVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDQNZWGAYDQMBVGE> > . > You are receiving this because you were mentioned.Message ID: > ***@***.***> >

kengoon · 2025-05-15T06:35:05Z

@T-Dynamos I think same should be done for sdl3_image

AndreMiras · 2025-05-15T07:08:35Z

pythonforandroid/recipes/sdl2_image/__init__.py

+        build_dir = self.get_build_dir(arch.arch)
+
+        with open(os.path.join(build_dir, ".gitmodules"), "r") as file:
+            for section in file.read().split('[submodule "')[1:]:
+                line_split = section.split(" = ")
+                # Parse .gitmodule section
+                clone_path, url, branch = (
+                    os.path.join(build_dir, line_split[1].split("\n")[0].strip()),
+                    line_split[2].split("\n")[0].strip(),
+                    line_split[-1].strip()
+                )
+                # Clone if needed
+                if not os.path.exists(clone_path) or not os.listdir(clone_path):
+                    shprint(
+                        sh.git, "clone", url,
+                        "--depth", "1", "-b",
+                        branch, clone_path, "--recursive"
+                    )


I should have commented that before, but when it's not that obvious what the code is doing, a comment is always helpful.
Maybe we could have added a block comment or method docstring saying:

# Prevents downloading commit history of all submodules, to improve cloning reliability on slow internet connection

AndreMiras · 2025-05-15T07:10:24Z

Thanks @AndiEcker for testing this too and thank you @T-Dynamos for making the fix 🙏

sdl2_image: fix downloading

d9ded99

kuzeyron added the recipe label Apr 10, 2025

clayote and others added 3 commits May 13, 2025 04:47

Merge branch 'develop' into sdl2_image_fix

dc9a1d6

Fix typo in comment

5a844e7

Remove redundant file.close()

2d5ef7f

The end of a `with` block already does that.

Fix bad empty-directory check

d2911a7

clayote approved these changes May 14, 2025

View reviewed changes

clayote merged commit be3de2e into kivy:develop May 14, 2025
30 of 33 checks passed

AndreMiras reviewed May 15, 2025

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

`sdl2_image`: fix downloading #3135

`sdl2_image`: fix downloading #3135

Uh oh!

T-Dynamos commented Mar 26, 2025

Uh oh!

misl6 commented Apr 8, 2025

Uh oh!

T-Dynamos commented Apr 9, 2025

Uh oh!

AndiEcker commented Apr 9, 2025 via email

Uh oh!

T-Dynamos commented May 12, 2025

Uh oh!

clayote commented May 12, 2025

Uh oh!

clayote commented May 12, 2025

Uh oh!

AndiEcker commented May 13, 2025 via email

Uh oh!

T-Dynamos commented May 13, 2025 •

edited

Loading

Uh oh!

AndiEcker commented May 13, 2025 via email

Uh oh!

AndiEcker commented May 14, 2025 via email

Uh oh!

Uh oh!

kengoon commented May 15, 2025

Uh oh!

AndreMiras May 15, 2025

Uh oh!

AndreMiras commented May 15, 2025

Uh oh!

Uh oh!

sdl2_image: fix downloading #3135

sdl2_image: fix downloading #3135

Uh oh!

Conversation

T-Dynamos commented Mar 26, 2025

Uh oh!

misl6 commented Apr 8, 2025

Uh oh!

T-Dynamos commented Apr 9, 2025

Uh oh!

AndiEcker commented Apr 9, 2025 via email

Uh oh!

T-Dynamos commented May 12, 2025

Uh oh!

clayote commented May 12, 2025

Uh oh!

clayote commented May 12, 2025

Uh oh!

AndiEcker commented May 13, 2025 via email

Uh oh!

T-Dynamos commented May 13, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

AndiEcker commented May 13, 2025 via email

Uh oh!

AndiEcker commented May 14, 2025 via email

Uh oh!

Uh oh!

kengoon commented May 15, 2025

Uh oh!

AndreMiras May 15, 2025

Choose a reason for hiding this comment

Uh oh!

AndreMiras commented May 15, 2025

Uh oh!

Uh oh!

`sdl2_image`: fix downloading #3135

`sdl2_image`: fix downloading #3135

T-Dynamos commented May 13, 2025 •

edited

Loading