idf.py build progress showing percentage instead of line numbers

[mbratch]

I have esp-idf 5.5.3 installed on two different machines. I might have installed them using the same procedure, I don't recall. But I did not set any special options or anything.

I've been through a lot of esp-idf upgrades and installs, and when I do an "idf.py build" it has always shown the progress as line number ([current-line / total-lines]). However, one of my two current installations has a different looking output. It's a bit more verbose but not nearly as verbose as using the -v option. Also, it shows progress as [XX %] instead of line numbers.

This is no big deal but I have a soft preference for the line numbers. Anyone know how to control what this output looks like? I read through the doc page for idf.py but didn't see anything about it.

[nopnop2002]

Is it possible that the versions of ESP-IDF installed on the two PCs are different?
You can check the ESP-IDF version with this command.

Code: Select all

$ idf.py --version
ESP-IDF v5.5.3-dirty

[mbratch]

Is it possible that the versions of ESP-IDF installed on the two PCs are different?
You can check the ESP-IDF version with this command.

Code: Select all

$ idf.py --version
ESP-IDF v5.5.3-dirty

Good question! They both show ESP-IDF v5.5.3.
It seems like there must be some kind of option or something that drives this format of output. I've been through many ESP-IDF versions over the last few years and have never seen this particular build output format;

Code: Select all

[ 31%] Building C object esp-idf/vfs/CMakeFiles/__idf_vfs.dir/vfs_eventfd.c.obj
[ 31%] Building C object esp-idf/vfs/CMakeFiles/__idf_vfs.dir/vfs_semihost.c.obj
[ 32%] Building C object esp-idf/vfs/CMakeFiles/__idf_vfs.dir/nullfs.c.obj
[ 32%] Linking C static library libvfs.a

[nopnop2002]

I don't know what caused it, but it's very interesting.
Is it possible that the versions of ninja installed on the two PCs are different?

Code: Select all

$ ninja --version
1.11.1

[mbratch]

I don't know what caused it, but it's very interesting.
Is it possible that the versions of ninja installed on the two PCs are different?

Code: Select all

$ ninja --version
1.11.1

Is it possible that my environment uses an alternative to ninja? It is not in one of my .espressif installations. This is the one that does percentages. This environment is Ubuntu running in a VirtualBox.

My other environment, which shows line numbers, is running ninja 1.12.1. It's native Ubuntu.

[nopnop2002]

Is it possible that my environment uses an alternative to ninja?

Maybe so.

In environments where ninja is installed, idf.py use ninja.
`idf.py build` will display this:

Code: Select all

Executing "ninja all"...
[4/1085] Generating ../../partition_table/partition-table.bin

In environments where ninja is not installed, idf.py use make.
`idf.py build` will display this:

Code: Select all

Executing "make -j 6 all"...
[  0%] Generating ../../partition_table/partition-table.bin
[  1%] Generating project_elf_src_esp32.c
[  1%] Generating /home/nop/rtos/version/build/esp-idf/esp_system/ld/sections.ld.in linker script...
[  1%] Generating /home/nop/rtos/version/build/esp-idf/esp_system/ld/memory.ld linker script...
[  1%] Built target custom_bundle
[  1%] Built target _project_elf_src
[  1%] Built target sections.ld.in

My other environment, which shows line numbers, is running ninja 1.12.1. It's native Ubuntu.

Try this on native Ubuntu.

Code: Select all

$ sudo apt remove --purge ninja-build
$ idf.py fullclean
$ idf.py build

Executing "make -j 6 all"...

$ sudo apt install ninja-build
$ idf.py fullclean
$ idf.py build

Executing "ninja all"...

[mbratch]

Thanks @nopnop2002. What you describe explains what I'm seeing. I don't recall doing anything but a "default" procedure installing ESP-IDF on either platform, but on the virtual box, it must have gone down the "make" path. It's functioning properly, so I may leave it be ("let a sleeping dog lie..." "if it ain't broke, don't fix it..." etc )