Hi Hamidreza,
You’re initiated a great thread. Reading all the responses makes me think it
would be awesome if this list had an archivist. There is so much useful
information spread around the various people on the list, someone that could
collect and publish it would provide a great service. End of tangent.
Regarding your issue, I love the python repl. It probably doesn’t apply so
well for getting HTML pages full of JavaScript, but if you’re hitting RESTful
APIs with JSON payloads, I’ve found it very helpful.
I’ve been poking around the Atlassian Jira APIs using the repl to learn how the
API really works, and discover some of the differences between, for example the
server and agile APIs. I’ve built up helper functions in a module to extract
the JSON, create an object, look into that object at various levels including
understanding the types of the values of the properties. I mostly use the
requests and json modules.
Regarding how to know what the payload should be, that’s harder if you can’t
hit it easily with your browser. You mentioned they are Java based
micro-services, do you have access to the Java source code? Looking at the
code might help clarify what the payload should be.
Can the micro-services be isolated and used independently? Do you know if any
of them maintain state? If they can be isolated, maybe you can hit the one you
want with a request. Though keep in mind there might still be coupling between
the services, e.g., service X queries the authentication service to check the
validity of a passed in token.
When I hear 60 micro-services, I automatically think many teams with a probably
kludged together architecture. If this is a platform new to you, maybe it
would help to meet with a senior architect to get an overview. At least try to
understand who is responsible for which components.
Also, if you’re looking at a specific bug, it’s likely not all 60 services are
involved in that transaction. If you haven’t already, understanding the many
workflows and which micro-services are part of which flows is very helpful.
There might be (probably inaccessible) documents (and probably outdated)
describing some of that information. Again that might be something to ask an
architect or technical lead about.
It sounds from your initial post like you’ve been thrown in to a large project
with little guidance. Understanding the structure and people in the
organization that has created that backend can be very helpful in understanding
the backend itself. And identifying who on the team is most helpful and best
at describing problems and where to start looking will be very helpful in the
long run.
Hope that helps.
Joel
From: program-l-bounce@xxxxxxxxxxxxx <program-l-bounce@xxxxxxxxxxxxx> On Behalf
Of hamidreza abroshan
Sent: Monday, January 17, 2022 12:40 PM
To: program-l@xxxxxxxxxxxxx
Subject: [program-l] Re: If you had to work on a web project with lots of
microservices that does not have an accessible ui, How do you find which part
of the code needs change or has bug?
Thanks Timothy for your brilliant solutions!
About screen reader, I'm using nvda too.
Accessibility is not a priority of this company and I'm not familiar with the
front end very much, and I only know a little bit of html.
But the bigger problem is that I was hired as a back end developer and If I
want to do these things, I need more time and should create a task in jira for
that.
Are you working as a back end developer?
How did you convince them to put more time on accessibility?
I would be glad if you provide info that helps me make the app more WCAG
compliant.
About http requests, I'm using firefox .
I think you are familiar with copying requests within firefox, but I explain
that it may be useful for someone.
As you know, pressing F12 opens Firefox dev tools. and will track every request
and response from now on.
Here, if you press Nvda+Space, you can easily get Http requests info.
if you press ctrl+home, you will find some buttons regarding various tabs of
the dev tools and the network is selected by default.
By pressing arrow down key, you can find some buttons to filters some requests
like images, html pages and so on and by pressing t, you can finde table that
shows list of requests and responses.
If you press enter on any Http request and press applications key and open
coppy sub menu, you can coppy request, request headers, post data, response
headers and so on.
If you press enter on any Http request and press ctrl+end, You can read it's
details here. Reading these info is a little bit harder than copying them.
If your data is json, press on raw data and hear press ctrl+end and then space.
You can easily navigate json content.
But for reading headers, You should read them from the treeview above the json
editor by navigation mode.
I highly recommend to copy the data because it is much simpler.
Now, it's ready to be pasted on any rest client like advance rest client.
I'm recommend to use Advanced rest client instead of scripting.
I can say it is fully accessible.
Although scripting is fun, it is time consuming.
Thanks Timothy for your time and sharing your great experiences.
Best.
On Tue, Jan 18, 2022 at 12:02 AM hamidreza abroshan
<hamidreza.abroshan@xxxxxxxxx <mailto:hamidreza.abroshan@xxxxxxxxx> > wrote:
Travis,
The idea of macro is amazing!
I'm familiar with selenium and will search more about that and if you have more
info, I'll be very glad to hear from you.
Thanks!
On Mon, Jan 17, 2022 at 11:59 PM hamidreza abroshan
<hamidreza.abroshan@xxxxxxxxx <mailto:hamidreza.abroshan@xxxxxxxxx> > wrote:
Thanks all, these are very helpful recommendations and I 'll try to use all of
them as much as possible.
I don't know how to Thank you! You are the best friends that someone can find
in a mailing list!
I'll reply to all emails by sending an order.
On Mon, Jan 17, 2022 at 10:08 PM Dzhovani Chemishanov
<dzhovani.chemishanov@xxxxxxxxx <mailto:dzhovani.chemishanov@xxxxxxxxx> > wrote:
Hi,
my idea is not to run it in debug mode, but to have logs at debug
level. Or whatever way you can grap the production logs. Debug logs
are something totally different from running the application with
debug breakpoints.
If you don't have access to the production logs, you can hope that
either you can find the integration tests of the microservice
responsible and get an example from there or that you can reconstruct
the expected payload from the types of the variables in the handler.
If none of these works, you can try to reverse engineer the source
code of the ui and try to find what data could be put in the http
request.
Finally, you can ask a colleague to give you a hand with the payload.
HTH,
Dzhovani
On 1/17/22, timothyjb310@xxxxxxxxx <mailto:timothyjb310@xxxxxxxxx>
<timothyjb310@xxxxxxxxx <mailto:timothyjb310@xxxxxxxxx> > wrote:
Hi, so I work on a project with only a couple micro services, but have faced** To leave the list, click on the immediately-following link:-
similar problems. I use Google Chrome, and have slowly started to understand
how to use the dev console to help in debugging. I can provide what I know
and have done to read HTTP requests/responses, but like in Firefox, the
chrome console is not easy to use with a screen reader. I know as much as I
do from working with sited teammates. I use NVDA, and am basing all of my
instructions on using NVDA. JAWS may read things differently, and I can not
attest to doing any of this with JAWS.
In terms of working with an inaccessible UI, I understand. It is very
frustrating to be working on a story, but you cannot even test it until you
fix accessibility bugs. However, it does make the app more WCAG compliant,
and only makes your life easier. I tend to just use aria to improve
accessibility of the UI as I go. Eventually things get easier the more fixes
you make. I just slip in these accessibility fixes into the stories I work
on, and since I tend to only fix these bugs using aria, there is no visual
change to the screen so there is no complaints from product owners or anyone
else. My team, with my help, has become more aware of accessibility, and
things that can be done to write more accessible code. One of the big things
that I encourage my team to do, which tends to fix a lot of issues, is to
just use HTMl elements as they are ment to be used. So use the button tag
instead of a div or span tag marked up with a click event, or use a heading
tag instead of making the text bigger and bolding it. With aria, you can fix
these easily enough, but it is better to just use the actual elements as we
all know. If you have any questions about how to go about fixing these bugs
and others with aria, let me know.
If you are trying to examine HTTP requests and responses using Chrome, you
first need to open up the web page in question, then press f12. From the top
of the screen, down arrow until you find the network tab and press enter. It
is a tab like element not a link or button, so I have yet to find a quick
way to navigate the tabs. There may be chrome dev tool specific shortcuts
that may help. In newer versions of Chrome, you may need to down arrow to
the more tabs button and choose network from the dropdown, however my
version of Chrome does not have this. It is important that the network tab
is open before you test the feature that you are trying to read the HTTP
request/response for. If you are testing a request that is made when the
page loads, then just press f5 to refresh the page once the network tab is
open. To get back to your actual page, the fastest way I have found is to
jump to the address bar using alt+d, and shift tabbing until you start
hearing elements that are from your page. You should be able to navigate
your page as normal. Visually on the screen from what I have been told, the
dev tools window is still open, and is taking up part of the screen, where
the other part of the screen is the site. Navigate to the feature that is
causing you issues, and trigger the request. Once you know that the request
should be sent, do ald+d again to jump to your address bar, and tab into the
dev tools which should still be open. It gets a bit tricky from here, but I
recently figured out how to actually read the requests and responses,
although sometimes it takes some fiddling to get the information you want.
The quickest way I have found to get to the HTTP request/response tree, is
to first navigate to the end of the screen using control+end and then press
shift+x, which should read out the checkbox
“Network panel property page 3rd-party requests check box not checked
“
From hear, you should be able to tab three times, which should hopefully on
a good day turn on forms mode for you, place focus on the tree view, and
read out something like
“Network Log Rows: 11, use the up and down arrow keys to navigate and
interact with the rows of the table; Use browse mode to read cell by cell.”
You can use the arrow keys to navigate the tree view, if you want to examine
a request more carefully, expand that request. You can also press enter on
it, which should open the request, and give you some more options to view
headers and the response if you do insert+space to get out of forms mode. I
cannot instruct much beyond this point since I am still trying to find a
consistent method method to get to this information myself. Remember when
exploring to try right clicking on things to see what your options are using
shift+f10 or using your applications key.
Beyond getting the request from the UI, you may need to send it to an API.
As I am sure you know, postman is not accessible. It is a rather frustrating
thing for blind full stack developers, but there are alternatives. I
personally use httpi, which is a python based command line tool like curl,
but way easier to use. I have struggled until recently using httpi since my
company requires some additional flags and information to be used to
authenticate, but it works well enough. We use JWT authentication, so I
wrote a batch script using httpi and jq to request a token with one short
command. JQ is a command line tool also python based, which can parse json.
I just added this to my path. Then I can just run the command
Dev-token | clip
To request the token, and pipe it directly into my clipboard. Where
dev-token is the name of my batch file that I added to my path. I intend to
create other such batch files for our other environments, such as uat-token,
and prod-token, I have just not done it yet. This is assuming you are using
windows. Anyways, I save specific requests that I tend to make in batch
files that I can run later. I love scripting my development environment to
make my life easier, especially as a blind developer, since we tend to have
to do things differently and sometimes in a longer way than our sited peers.
A quick tip that I recently used with httpi is that if you assign your token
to the environment variable JWT_AUTH_TOKEN, and you use the jwt httpi
plugin, you can make requests without sending the token. If you want to know
more about this, let me know.
Hope this helped,
Timothy Breitenfeldt
From: program-l-bounce@xxxxxxxxxxxxx <mailto:program-l-bounce@xxxxxxxxxxxxx>
<program-l-bounce@xxxxxxxxxxxxx <mailto:program-l-bounce@xxxxxxxxxxxxx> > On
Behalf Of hamidreza abroshan
Sent: Monday, January 17, 2022 8:22 AM
To: program-l@xxxxxxxxxxxxx <mailto:program-l@xxxxxxxxxxxxx>
Cc: travis@xxxxxxxxxxxxxx <mailto:travis@xxxxxxxxxxxxxx>
Subject: [program-l] Re: If you had to work on a web project with lots of
microservices that does not have an accessible ui, How do you find which
part of the code needs change or has bug?
Devtools in firefox is accessible but a little bit tricky and you can
completely coppy http response, headers, and request and it's headers.
If you want, I can record quicke tutorial for that.
The trick is that you should use Nvda's focus mode instead of browse mode.
But my problem here is that The web UI is not accessible.
In some Dev companies, the test team is responsible for providing the
payload and should tell you which micro service has bugs, but here they
don't have such rules.
How can I automate with selenium if I can't browse some parts of the page?
Best.
On Mon, Jan 17, 2022 at 7:34 PM Karl-Otto Rosenqvist <karl-otto@xxxxxxxxxx
<mailto:karl-otto@xxxxxxxxxx>
<mailto:karl-otto@xxxxxxxxxx ;<mailto:karl-otto@xxxxxxxxxx> > > wrote:
Hi!
My experience is that the dev tools in Microsoft Edge is hard to use
with NVDA. Perhaps it's because I don't have enough experience and
knowledge of where to look.
I'm partly sighted and have tried to inspect elements but I feel that
it's very hard to do with a screen reader. Perhaps it is easier if you
know the keyboard shortcuts to jump between the dev tools window/pane
and the page itself?
Perhaps it's a similar problem that hamidreza abroshan experiences?
Regards
Karl-Otto
Karl-Otto Rosenqvist
MAWINGU
Orgnr: 750804-3937
0701- 75 98 56
karl-otto@xxxxxxxxxx <mailto:karl-otto@xxxxxxxxxx>
<mailto:karl-otto@xxxxxxxxxx ;<mailto:karl-otto@xxxxxxxxxx> >
https://mawingu.se
Den 2022-01-17 kl. 15:37, skrev travis@xxxxxxxxxxxxxx
<mailto:travis@xxxxxxxxxxxxxx>
<mailto:travis@xxxxxxxxxxxxxx ;<mailto:travis@xxxxxxxxxxxxxx> > :
SO are you saying the web application is inaccessible? Or the dev tools** To leave the list, click on the immediately-following link:-
are?
If the web application is inaccessible that seems a whole problem on its
own.
I guess if you really wanted you could try to automate executing steps
in the web application using something like Selenium. But it sure sounds
like an uphill battle if the application’s UI is inaccessible to that
extreme.
*From:* program-l-bounce@xxxxxxxxxxxxx
<mailto:program-l-bounce@xxxxxxxxxxxxx>
<mailto:program-l-bounce@xxxxxxxxxxxxx ;
<mailto:program-l-bounce@xxxxxxxxxxxxx> > <program-l-bounce@xxxxxxxxxxxxx
<mailto:program-l-bounce@xxxxxxxxxxxxx>
<mailto:program-l-bounce@xxxxxxxxxxxxx ;
<mailto:program-l-bounce@xxxxxxxxxxxxx> > >
*On Behalf Of *hamidreza abroshan
*Sent:* Monday, January 17, 2022 7:12 AM
*To:* program-l@xxxxxxxxxxxxx <mailto:program-l@xxxxxxxxxxxxx>
<mailto:program-l@xxxxxxxxxxxxx ;<mailto:program-l@xxxxxxxxxxxxx> >
*Subject:* [program-l] Re: If you had to work on a web project with lots
of microservices that does not have an accessible ui, How do you find
which part of the code needs change or has bug?
Thanks Dzhovani,
We are running micro services locally, and can not trigger http requests
from outside.
If I could debug the Micro services that are running on the central
server, the test team and other developers can not use the app because I
put a breakpoint somewhere and the execution stops when it hits my break
point.
Let me explain my problem clearly.
Imagine that you joined a back end team of an app.
You receive a bug that says, "in the second page of customer
registration, when customer details are added, We get 500 server error".
At this point, your fellow sited developer, opens the UI, navigates to
registration, completes the form until gets to page 2, and in page 2,
before hitting that button that sends customer details, opens his
browser dev tools to track http requests and hits that send button.
Now, he can copy the http request and response from dev tools and resend
with any rest client like postman and debug the local app, but if you
were blind, you can not sea the UI, and you can not obtain the payload
that you need to debug the app this way.
I'm asking any alternate solutions that
On Mon, Jan 17, 2022 at 1:39 PM Dzhovani Chemishanov
<dzhovani.chemishanov@xxxxxxxxx <mailto:dzhovani.chemishanov@xxxxxxxxx>
<mailto:dzhovani.chemishanov@xxxxxxxxx ;
<mailto:dzhovani.chemishanov@xxxxxxxxx> >
<mailto:dzhovani.chemishanov@xxxxxxxxx ;
<mailto:dzhovani.chemishanov@xxxxxxxxx>
<mailto:dzhovani.chemishanov@xxxxxxxxx ;
<mailto:dzhovani.chemishanov@xxxxxxxxx> > >>
wrote:
Hi,
As far as I get your problem, you are unable to trigger an http
request in order to inspect its content. If I get you correctly,
debug
log the incoming request on the backend and wait for someone to
trigger it. This is valid if you can fiddle with the handlers and
have access to the raw data of the http request.
HTH,
Dzhovani
On 1/17/22, hamidreza abroshan <hamidreza.abroshan@xxxxxxxxx
<mailto:hamidreza.abroshan@xxxxxxxxx>
<mailto:hamidreza.abroshan@xxxxxxxxx ;<mailto:hamidreza.abroshan@xxxxxxxxx> >
<mailto:hamidreza.abroshan@xxxxxxxxx ;
<mailto:hamidreza.abroshan@xxxxxxxxx>
<mailto:hamidreza.abroshan@xxxxxxxxx ;<mailto:hamidreza.abroshan@xxxxxxxxx> >
> Hey all,wrote:
> I'm working on a java project that has 60 microservices and this
project
> does not have an accessible UI.
> My problem is that when I need to debug this app, I need a
payload to send
> with rest clients, and every one says, "go and get the payload
from UI."
> What do you do in similar situations?
> I need urgent help and appreciate your help.
> Thanks in advance.
>
> --
> hamidreza
>
** To leave the list, click on the immediately-following link:-
** [mailto:program-l-request@xxxxxxxxxxxxx ;
<mailto:program-l-request@xxxxxxxxxxxxx>
<mailto:program-l-request@xxxxxxxxxxxxx ;
<mailto:program-l-request@xxxxxxxxxxxxx> >
<mailto:program-l-request@xxxxxxxxxxxxx ;
<mailto:program-l-request@xxxxxxxxxxxxx>
<mailto:program-l-request@xxxxxxxxxxxxx ;
<mailto:program-l-request@xxxxxxxxxxxxx> > >?subject=unsubscribe]
** If this link doesn't work then send a message to:
** program-l-request@xxxxxxxxxxxxx
<mailto:program-l-request@xxxxxxxxxxxxx>
<mailto:program-l-request@xxxxxxxxxxxxx ;
<mailto:program-l-request@xxxxxxxxxxxxx> >
<mailto:program-l-request@xxxxxxxxxxxxx ;
<mailto:program-l-request@xxxxxxxxxxxxx>
<mailto:program-l-request@xxxxxxxxxxxxx ;
<mailto:program-l-request@xxxxxxxxxxxxx> > >
** and in the Subject line type
** unsubscribe
** For other list commands such as vacation mode, click on the
** immediately-following link:-
** [mailto:program-l-request@xxxxxxxxxxxxx ;
<mailto:program-l-request@xxxxxxxxxxxxx>
<mailto:program-l-request@xxxxxxxxxxxxx ;
<mailto:program-l-request@xxxxxxxxxxxxx> >
<mailto:program-l-request@xxxxxxxxxxxxx ;
<mailto:program-l-request@xxxxxxxxxxxxx>
<mailto:program-l-request@xxxxxxxxxxxxx ;
<mailto:program-l-request@xxxxxxxxxxxxx> > >?subject=faq]
** or send a message, to
** program-l-request@xxxxxxxxxxxxx
<mailto:program-l-request@xxxxxxxxxxxxx>
<mailto:program-l-request@xxxxxxxxxxxxx ;
<mailto:program-l-request@xxxxxxxxxxxxx> >
<mailto:program-l-request@xxxxxxxxxxxxx ;
<mailto:program-l-request@xxxxxxxxxxxxx>
<mailto:program-l-request@xxxxxxxxxxxxx ;
<mailto:program-l-request@xxxxxxxxxxxxx> > > with the Subject:- faq
--
hamidreza
** [mailto:program-l-request@xxxxxxxxxxxxx ;
<mailto:program-l-request@xxxxxxxxxxxxx>
<mailto:program-l-request@xxxxxxxxxxxxx ;
<mailto:program-l-request@xxxxxxxxxxxxx> > ?subject=unsubscribe]
** If this link doesn't work then send a message to:
** program-l-request@xxxxxxxxxxxxx <mailto:program-l-request@xxxxxxxxxxxxx>
<mailto:program-l-request@xxxxxxxxxxxxx ;
<mailto:program-l-request@xxxxxxxxxxxxx> >
** and in the Subject line type
** unsubscribe
** For other list commands such as vacation mode, click on the
** immediately-following link:-
** [mailto:program-l-request@xxxxxxxxxxxxx ;
<mailto:program-l-request@xxxxxxxxxxxxx>
<mailto:program-l-request@xxxxxxxxxxxxx ;
<mailto:program-l-request@xxxxxxxxxxxxx> > ?subject=faq]
** or send a message, to
** program-l-request@xxxxxxxxxxxxx <mailto:program-l-request@xxxxxxxxxxxxx>
<mailto:program-l-request@xxxxxxxxxxxxx ;
<mailto:program-l-request@xxxxxxxxxxxxx> >
with the Subject:- faq
--
hamidreza
