Slow powershell startup #4811

@Fonta commented on GitHub (Mar 9, 2020):

I think I might the same thing here. By times when I start Windows Terminal i simply get a blinking cursor for around 7-10 seconds or so.
This happened when I had pwsh 5.1 as default and now also with pwsh 7 as default.
There's nothing in my profile that gets started.

The weird thing is also that it isn't always slow. A lot of times it's as fast as starting PS normally.

@Fonta commented on GitHub (Mar 9, 2020): I think I might the same thing here. By times when I start Windows Terminal i simply get a blinking cursor for around 7-10 seconds or so. This happened when I had pwsh 5.1 as default and now also with pwsh 7 as default. There's nothing in my profile that gets started. The weird thing is also that it isn't always slow. A lot of times it's as fast as starting PS normally.

claunia commented

@DHowett-MSFT commented on GitHub (Mar 10, 2020):

Contrasting running pwsh directly, which never takes a long time? I'm somewhat skeptical: we're really just spawning it in the exact same way as always...

@DHowett-MSFT commented on GitHub (Mar 10, 2020): Contrasting running pwsh directly, which _never_ takes a long time? I'm somewhat skeptical: we're really just spawning it in the exact same way as always...

claunia commented

@cdmihai commented on GitHub (Mar 10, 2020):

@DHowett-MSFT, maybe there's a measurement gap somewhere. My $Profile script prints out its time at the end, and Terminal (or something else) still does something 1-4 seconds after that, until I can type. Ideally it would be nice to measure 3 times;

terminal startup to shell invocation
shell startup time
time from end of shell startup to when user can type stuff

@cdmihai commented on GitHub (Mar 10, 2020): @DHowett-MSFT, maybe there's a measurement gap somewhere. My $Profile script prints out its time at the end, and Terminal (or something else) still does something 1-4 seconds after that, until I can type. Ideally it would be nice to measure 3 times; - terminal startup to shell invocation - shell startup time - time from end of shell startup to when user can type stuff

claunia commented

@DHowett-MSFT commented on GitHub (Mar 10, 2020):

Terminal (or something else) still does something 1-4 seconds after that, until I can type.

By this point, terminal is just two pipes hooked up to a console process; it cannot introduce further delay between when the shell prints something and accepts input.
Do you, per chance, have a complicated Prompt function?

Regardless, would you mind capturing traces using this WPR profile, using the Console Performance Default Profile? Thanks!

My e-mail (and therefore MS alias ;P) is in my github profile.

@DHowett-MSFT commented on GitHub (Mar 10, 2020): > Terminal (or something else) still does something 1-4 seconds after that, until I can type. By this point, terminal is just two pipes hooked up to a console process; it cannot introduce further delay between when the shell prints something and accepts input. Do you, per chance, have a complicated `Prompt` function? Regardless, would you mind capturing traces using [this WPR profile](https://raw.githubusercontent.com/microsoft/terminal/dev/duhowett/wprp/src/TerminalPerf.wprp), using the `Console Performance Default Profile`? Thanks! My e-mail (and therefore MS alias ;P) is in my github profile.

claunia commented

@cdmihai commented on GitHub (Mar 11, 2020):

Do you, per chance, have a complicated Prompt function?

I do have posh-git installed, but my startup directory is not in a git repo.

I'll get a recording and send it to you.

@cdmihai commented on GitHub (Mar 11, 2020): >Do you, per chance, have a complicated Prompt function? I do have posh-git installed, but my startup directory is not in a git repo. I'll get a recording and send it to you.

claunia commented

@ghost commented on GitHub (May 4, 2020):

This issue has been automatically marked as stale because it has been marked as requiring author feedback but has not had any activity for 4 days. It will be closed if no further activity occurs within 3 days of this comment.

@ghost commented on GitHub (May 4, 2020): This issue has been automatically marked as stale because it has been marked as requiring author feedback but has not had any activity for **4 days**. It will be closed if no further activity occurs **within 3 days of this comment**.

claunia commented

@DHowett-MSFT commented on GitHub (May 7, 2020):

@cdmihai guessing that posh-git ended up being the culprit?

@DHowett-MSFT commented on GitHub (May 7, 2020): @cdmihai guessing that posh-git ended up being the culprit?

claunia commented

@cdmihai commented on GitHub (May 7, 2020):

That's my current lead, per our email discussion. Unfortunately I didn't manage to carve out enough time to study it further, and posh-git is too useful to remove. But one of these days ...

@cdmihai commented on GitHub (May 7, 2020): That's my current lead, per our email discussion. Unfortunately I didn't manage to carve out enough time to study it further, and posh-git is too useful to remove. But one of these days ...

claunia commented

@archongum commented on GitHub (Oct 19, 2020):

That's my current lead, per our email discussion. Unfortunately I didn't manage to carve out enough time to study it further, and posh-git is too useful to remove. But one of these days ...

Same here. I need post-git and startup speed too.

@archongum commented on GitHub (Oct 19, 2020): > That's my current lead, per our email discussion. Unfortunately I didn't manage to carve out enough time to study it further, and posh-git is too useful to remove. But one of these days ... Same here. I need `post-git` and startup speed too.

claunia commented

@DHowett commented on GitHub (Oct 19, 2020):

It would be worth filing a bug on them to get git.exe out of the critical path then!

@DHowett commented on GitHub (Oct 19, 2020): It would be worth filing a bug on them to get git.exe out of the critical path then!

claunia commented

@danicunhac commented on GitHub (Oct 29, 2020):

I have the same problem :/ Seems like posh really is causing the slow down

@danicunhac commented on GitHub (Oct 29, 2020): I have the same problem :/ Seems like posh really is causing the slow down

claunia commented

Link to stackoverflow solution

@danicunhac commented on GitHub (Oct 29, 2020):

@danicunhac commented on GitHub (Oct 29, 2020): [Link to stackoverflow solution](https://stackoverflow.com/questions/48262269/how-to-i-remove-the-powershell-start-text) ![image](https://user-images.githubusercontent.com/28742636/97610576-d7035800-19f3-11eb-9429-077244c1323c.png)

claunia commented

@DHowett commented on GitHub (Oct 29, 2020):

I do not know why you have decided that powershell printing 4 lines of text is at all related to its speed.

@DHowett commented on GitHub (Oct 29, 2020): I do not know why you have decided that powershell printing 4 lines of text is at all related to its _speed_.

claunia commented

@cdmihai commented on GitHub (Oct 29, 2020):

I do not know why you have decided that powershell printing 4 lines of text is at all related to its speed.

I decided to give it a try because why not, and it is indeed a major improvement in startup speed. Seems ridiculous, but alas that's software for ya :)

@cdmihai commented on GitHub (Oct 29, 2020): > I do not know why you have decided that powershell printing 4 lines of text is at all related to its speed. I decided to give it a try because why not, and it is indeed a major improvement in startup speed. Seems ridiculous, but alas that's software for ya :)

claunia commented

@danicunhac commented on GitHub (Oct 29, 2020):

I didn't decide, it was just a coincidence. I wanted to remove that 4 lines of text and it also improved startup speed (also looks better noew). So yeah :)

@danicunhac commented on GitHub (Oct 29, 2020): I didn't decide, it was just a coincidence. I wanted to remove that 4 lines of text and it also improved startup speed (also looks better noew). So yeah :)

claunia commented

@DHowett commented on GitHub (Oct 29, 2020):

It's 2020. Weirder things have happened! 😄

@DHowett commented on GitHub (Oct 29, 2020): It's 2020. Weirder things have happened! :smile:

claunia commented